In a groundbreaking development for the aviation industry, researchers have unveiled an innovative intelligent fault diagnosis method designed specifically for general aviation aircraft. This new framework, detailed in a recent paper on arXiv, leverages a multi-fidelity digital twin approach combined with advanced Failure Mode and Effects Analysis (FMEA) to tackle the persistent challenges of fault diagnosis.

General aviation aircraft often face significant hurdles due to the scarcity of real fault data, the diversity of fault types, and the weak signatures that accompany these faults. The proposed solution integrates four key modules: high-fidelity flight dynamics simulation, FMEA-driven fault injection, multi-fidelity residual feature extraction, and a large language model (LLM)-enhanced report generation system.

At the core of this framework is a digital twin constructed using the JSBSim six-degree-of-freedom (6-DoF) flight dynamics engine, which generates comprehensive engine health monitoring data. A sophisticated three-layer fault injection engine models the physical causal propagation of 19 distinct engine fault types, ensuring a thorough analysis of potential issues.

The multi-fidelity residual computation framework employs paired-mirror residuals and a GRU surrogate prediction model to enhance diagnostic accuracy. Notably, the high-fidelity path captures clean fault deviation signals, while the low-fidelity path enables real-time residual computation. An end-to-end diagnosis is performed using a 1D-CNN classifier that accurately identifies 20 fault classes.

Impressively, experiments have shown that the paired-mirror residual scheme achieves a Macro-F1 score of 96.2%, while the GRU surrogate model accelerates inference by 4.3 times with minimal performance cost. This research establishes a new design principle, emphasizing that the quality of residual features is paramount to diagnostic performance, outperforming traditional classifier architecture by a factor of five.

This advancement not only promises to enhance safety in general aviation but also sets a new standard for fault diagnosis across the aerospace industry.